A room temperature control method based on dynamic room temperature set value and model predictive control

Abstract

The present application belongs to the technical field of heating room temperature control, and discloses a room temperature control method based on dynamic room temperature setting value and model predictive control. The present application relates to a multi-objective optimization model taking hourly heat load as energy consumption and PMV-PPD as indoor thermal comfort, and uses a genetic algorithm to solve room temperature setting values in different time periods within a day. The method can formulate an energy-saving and comfortable room temperature setting value scheme in advance according to weather forecasts. The model predictive control method is used to regulate and control the heating room temperature, so as to ensure that the room temperature changes according to the dynamic room temperature setting value trajectory. The method of the present application technically formulates an energy-saving and comfortable room temperature setting value scheme in advance according to weather forecasts on the user side of the heating system, and realizes the room temperature control effect of comfort and energy saving in application, so that the central heating system can continuously realize large-scale room temperature setting value optimization.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of heat supply room temperature control, and particularly relates to a room temperature control method based on dynamic room temperature setting values and model predictive control. BACKGROUND

[0002] The central heating system is one of the urban infrastructures and one of the modernization level symbols of the city. With the continuous increase of the heating area, the problems of poor thermal comfort and energy waste caused by excessively high indoor temperature are paid more and more attention. Properly reducing the indoor temperature setting value can reduce the energy consumption and keep the thermal comfort of indoor personnel within the allowable range.

[0003] Many heating enterprises usually design and calculate according to the maximum value of the indoor temperature setting value in the specification in order to ensure that the indoor temperature of the heat user reaches the requirement, but this is not energy-saving. With the change of the weather, the dynamic room temperature setting value for the future one day is formulated in advance, which can ensure the thermal comfort and reduce the heating energy consumption. However, how to make the heating system automatically formulate the dynamic room temperature setting value for the future one day according to the weather forecast and ensure that the thermal comfort of the indoor heat user is within the allowable range becomes a key problem for the implementation of the dynamic room temperature setting value strategy. At the same time, how to ensure that the room temperature changes according to the trajectory of the dynamic room temperature setting value also becomes a key problem for the implementation of the dynamic room temperature setting value. SUMMARY

[0004] In order to solve the above problems, the application particularly provides a room temperature control method based on dynamic room temperature setting values and model predictive control, involves a multi-objective optimization model taking the hourly heat load as the energy consumption and taking the PMV-PPD as the indoor thermal comfort, and uses the genetic algorithm to solve the room temperature setting values in different time periods within one day. The method can formulate an energy-saving and comfortable room temperature setting value scheme in advance according to the weather forecast. The model predictive control method is adopted to regulate and control the heating room temperature, so that the room temperature changes according to the trajectory of the dynamic room temperature setting value.

[0005] To achieve the above purpose, the technical scheme of the application is as follows:

[0006] A room temperature control method based on dynamic room temperature setting values and model predictive control, the steps are as follows:

[0007] Step S1, determine the heating building or heating room, and establish an hourly heat load prediction model of the heating building or heating room;

[0008] The hourly heat load prediction model selects a neural network prediction model, and the training data is obtained by an energy consumption simulation software or collected actual operation data. In order to improve the sensitivity of the hourly heat load prediction model to the room temperature setting value, the room temperature setting value in the training data needs to change in 18-24℃.

[0009] The input variables of the hourly heat load prediction model are: indoor design temperature; whether the current time is working time; outdoor dry bulb temperature; solar radiation; occupancy rate; light usage rate; heat load values of 1 h, 2 h, 24 h before the current time;

[0010] The output variable of the hourly heat load prediction model is the hourly heat load value of the next time;

[0011] Step S2, using the hourly heat load prediction model and the simplified PPD calculation program to constitute a multi-objective optimization model, and determining the solving method;

[0012] Among them, the variable reasonable simplification method in the PPD calculation program is: the mechanical work done by the human body is 0, the basic thermal resistance of clothing is 1.4 clo, the human body metabolism rate is 1 Met, the average radiation temperature is 2℃ lower than the room temperature, the indoor air humidity is the water vapor partial pressure under the room temperature Tn and the relative humidity of 40%, and the indoor air flow rate v is 0.07 m / s; the simplified PPD calculation program is denoted as:

[0013] (1)

[0014] The multi-objective optimization model constituted by the hourly heat load prediction model and the simplified PPD calculation program is,

[0015] (2)

[0016] In formula (2), is the room heat load index, is the room thermal comfort index; is the sum of the heat load prediction values under the dynamic room temperature setting value within 24 hours, that is, the energy consumption prediction value of the whole day, kWh; is the hourly heat load prediction model in step S1, in actual use, The calculation of needs to input all the variables required by the hourly heat load prediction model mentioned in step S1, wherein the meteorological parameters are obtained from the weather forecast; is the average predicted percentage of dissatisfaction of the personnel in a day, calculated by formula (1);

[0017] Among them, the decision variable of the multi-objective optimization model is the room temperature setting value, and it is stipulated that the room temperature setting value does not change within 3 hours, the value range is {18℃, 19℃, 20℃, 21℃, 22℃, 23℃, 24℃}, and the solving method is genetic algorithm for solving;

[0018] Step S3, design and add model predictive controller for heating building or heating room;

[0019] Add one model predictive controller for each building or for each heat user, add according to heat user demand;

[0020] The prediction model selects white box model, black box model or gray box model, but needs to meet the following functions: input current room temperature, current meteorological condition and radiator hot water flow, output next time indoor temperature;

[0021] The target of the rolling optimization of the model predictive controller is to track and optimize the room temperature set value obtained with the minimum room temperature fluctuation and water pump delivery energy consumption, and the objective function is expressed as,

[0022] (3)

[0023] In the formula, N is the prediction time domain, and k =0 is defined as the sampling time, and each rolling optimization starts from k =0, y k is the room temperature prediction value of the k th time step, y 0 is the real value of the measured room temperature, P k is the variable frequency water pump energy consumption in the k th time step in the prediction time domain; is the weight coefficient of the room temperature fluctuation and the variable frequency water pump energy consumption;

[0024] The hot water flow sequence in the prediction time domain in the rolling optimization of the model predictive controller is expressed by U , and the optimization range of each component u is

[0025] (4)

[0026] In the formula, is the maximum flow allowed to pass through the radiator;

[0027] The constraint condition in the rolling optimization of the model predictive controller is:

[0028] (5)

[0029] (6)

[0030] In the formula, y k is the room temperature set value of the kThe room temperature at any time should be greater than 18 DEG C specified in the regulation; a room temperature set value for the time, a deviation allowing fluctuation;

[0031] The internal logic of the model predictive controller is a rolling optimization process between the solution of the genetic algorithm and the sampling time; the genetic algorithm is mainly used to solve the optimal hot water flow sequence at each sampling time: firstly, different hot water flow sequences are randomly generated through the SEGA algorithm, then the real room temperature and the weather forecast are input into the prediction model, and the prediction model is used to obtain the room temperature prediction trajectory corresponding to the hot water flow sequence; then the room temperature prediction trajectory and the hot water flow are brought into the objective function in the rolling optimization to calculate the fitness, until the optimal hot water flow sequence and the cost-optimal room temperature prediction trajectory corresponding thereto are found; finally, the water pump frequency is adjusted to make the hot water flow through the radiator equal to the first item of the optimal hot water flow sequence, and the solution at the time is completed; the heating building or the heating room changes the room temperature through the radiator for a period of time, and at the next sampling time, the real room temperature is input into the prediction model, and the genetic algorithm is used to solve again, and the rolling optimization is completed.

[0032] Step S4, the user inputs the proportion of energy consumption and thermal comfort index according to the preference into the SEGA optimization model to obtain the optimal room temperature set value trajectory T;

[0033] The user's preference for the two objective functions will finally determine the selection of the decision variable, in order to obtain the final decision variable according to the user's preference, the objective functions are normalized and then weighted:

[0034] (7)

[0035] Wherein, i is 1 and 2, 、 are energy consumption index and thermal comfort index respectively, is the weight of the corresponding objective function set by the user;

[0036] Step S5, the model predictive controller controls the indoor temperature of the heating building or the heating room to change according to the optimal room temperature set value trajectory T.

[0037] Further, the room temperature control method of the present application needs to be calculated between 20:00 at night and 8:00 the next day every day, and the calculation content is the optimal room temperature set value trajectory T of the period from 8:00 the next day to 20:00 the day after tomorrow.

[0038] The beneficial effects of the present application: the present application specifically proposes an operation step and implementation scheme of a room temperature control method based on dynamic room temperature setting value and model predictive control, which technically completes the preparation of energy-saving and comfortable room temperature setting value scheme for the heat user side of the heating system according to the weather forecast, and realizes the room temperature control effect of comfort and energy saving in application, so that the central heating system can continuously realize the optimization of a wide range of room temperature setting values. First of all, the implementation of the method can reduce the heating energy consumption by reasonably reducing the room temperature setting value in a certain period of time; then, the model predictive controller design method proposed by the present application can match the dynamic room temperature setting value strategy proposed by the present application, so that the room temperature can change according to the dynamic room temperature setting value; finally, the room temperature control method based on dynamic room temperature setting value and model predictive control proposed by the present application can further reduce the energy consumption, realize energy saving and emission reduction, and lay a foundation for the formulation and implementation of the dynamic room temperature setting value of the heating user. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 It is a schematic diagram for establishing and operating the room temperature control method based on dynamic room temperature setting value and model predictive control of the present application.

[0040] Figure 2 It is a prediction effect diagram of the hourly heat load prediction model of an example.

[0041] Figure 3 It is an internal logic block diagram of the model predictive controller of an example.

[0042] In the figure: 1-room temperature setting value; 2-real value of room temperature; 3-weather forecast; 4-SEGA genetic algorithm; 5-room temperature prediction model; 6-hot water flow sequence; 7-weather data; 8-constraint condition; 9-objective function; 10-hot water flow optimization range; 11-room temperature sequence; 12-optimal hot water flow sequence; 13-cost-optimal room temperature prediction trajectory; 14-first component of the optimal hot water flow sequence; 15-corresponding frequency converter frequency;

[0043] Figure 4 It is a schematic diagram of the rolling optimization process of the model predictive controller of an example.

[0044] Fig. 1: 1 - current time; 2 - next time; 3 - hot water flow sequence generated at current time; 4 - actual room temperature at current time; 5 - weather forecast at current time; 6 - SEGA genetic algorithm; 7 - room temperature prediction model; 8 - constraint condition formed by room temperature set value at current time; 9 - objective function; 10 - hot water flow optimization range; 11 - room temperature sequence at current time; 12 - optimal hot water flow sequence at current time; 13 - room temperature prediction trajectory with optimal cost at current time; 14 - frequency of frequency converter corresponding to first component of optimal hot water flow sequence at current time; 15 - variable frequency water pump for controlling radiator flow; 16 - hot water flow sequence generated at next time; 17 - actual room temperature at next time; 18 - weather forecast at next time; 19 - constraint condition formed by room temperature set value at next time; 20 - room temperature sequence at next time; 21 - optimal hot water flow sequence at next time; 22 - room temperature prediction trajectory with optimal cost at next time; 23 - frequency of frequency converter corresponding to first component of optimal hot water flow sequence at next time;

[0045] Figure 5 Fig. 4: Room temperature tracking effect of model predictive control and traditional PID control under dynamic room temperature set value for an example simulation model.

[0046] Figure 6 Fig. 5: Room temperature control effect under different energy consumption and thermal comfort ratios for an example under the present application.

[0047] Figure 7 Fig. 6: Control effect under the same room temperature level for an example under the present application and traditional strategy.

[0048] Figure 8 Fig. 7: Heating heat consumption comparison under the present application and traditional strategy for an example.

[0049] Figure 9 Fig. 8: Heating pump consumption comparison under the present application and traditional strategy for an example. DETAILED DESCRIPTION

[0050] The technical solutions in the examples of the present application will be described clearly and completely below with reference to the drawings in the examples of the present application. Obviously, the described examples are only a part of the examples of the present application, rather than all the examples of the present application. Based on the examples in the present application, all other examples obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. It should be noted that the examples of the present application use simulation to verify the gain effect of the present application, and experiment can also be used to verify the gain effect of the present application. If the examples of the present application involve directional indications (such as up, down, left, right, east, south, west, north, etc.), the directional indications are only used to explain the relative positional relationship between rooms in the examples, and if the present application is applied to other buildings, the relative positional relationship between rooms should be determined according to specific circumstances.

[0051] In addition, if the description of "first", "second" and the like is involved in the examples of the present application, the description of "first", "second" and the like is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the technical solutions of each example can be combined with each other, but it must be based on the realization of the ordinary skilled in the art, when the combination of the technical solutions appears contradictory or unachievable, it should be considered that the combination of the technical solutions does not exist, and is not within the protection scope required by the present application.

[0052] The present application proposes a room temperature control method based on dynamic room temperature setting value and model predictive control, which aims to formulate a dynamic room temperature setting value according to the weather, and to make the room temperature change as much as possible according to the setting value trajectory, reduce energy consumption, realize energy saving and emission reduction, and bring economic benefits to the heating enterprise.

[0053] Reference Figure 1 In an example of the present application, a room temperature control method based on dynamic room temperature setting value and model predictive control comprises the following steps:

[0054] Step S1, determine the heating building or heating room, and establish the hourly heat load prediction model of the heating building or heating room;

[0055] Step S2, use the hourly heat load prediction model and the simplified PPD calculation program to constitute a multi-objective optimization model, and determine the solving method;

[0056] Step S3, design and add a model predictive controller for the heating building or room;

[0057] Step S4, the user inputs the proportion of energy consumption and thermal comfort index according to the preference to the SEGA optimization model, and obtains the optimal room temperature setting value trajectory T

[0058] Step S5, the model predictive controller controls the indoor temperature of the heating building or heating room to change according to the optimal room temperature setting value trajectory T;

[0059] In step S1 of the present example, the heating building is determined as a heating experimental room in Dalian City, one external window is arranged on each of the north-south external walls of the experimental room, an external door is arranged on the north external wall, the length, width and height of the room are 5m, 4m and 3m respectively, the heating equipment in winter is a common radiator, the room is used from 8:00 to 20:00, the fixed room temperature setting value is 18℃ during the non-use time, and there is no holiday;

[0060] The input variables for the hourly heat load prediction model in step S1 of this example are: "indoor design temperature", "whether the current time is working time", "outdoor dry-bulb temperature", "solar radiation", "occupancy rate", "lighting utilization rate", and "heat load values ​​of the previous 1 hour, 2 hours, and 24 hours". The output variable of the hourly heat load prediction model is the hourly heat load value for the next moment; for ease of description, it is denoted as...

[0061] (8)

[0062] In the formula, Set the room temperature value. To determine whether the current time is working hours. Outdoor dry-bulb temperature, This refers to solar radiation. For the staff occupancy rate, To maximize the utilization of lighting, This represents the heat load value for the previous hour at the current time. This represents the heat load value for the previous 2 hours at the current time. This represents the heat load value for the previous 24 hours at the current moment. The prediction results of the hourly heat load prediction model are as follows: Figure 2 As shown.

[0063] This example follows step S2 to construct a multi-objective optimization model.

[0064] Among them, the reasonable simplification method for variables in the PPD calculation program is to consider the mechanical work performed by the human body. The basic thermal resistance of the clothing is 0. The human metabolic rate is 1.4 clots. 1 Met, mean radiation temperature To be 2°C below room temperature, indoor air humidity Let Tn be the water vapor partial pressure at room temperature and 40% relative humidity, and v be the indoor air velocity at 0.07 m / s. The simplified PPD calculation procedure is denoted as follows:

[0065] (9)

[0066] The hourly heat load prediction model and the simplified PPD calculation program constitute a multi-objective optimization model.

[0067] (10)

[0068] In equation (10), This refers to the room's heat load index. Indicators for room thermal comfort; It is the sum of the predicted heat load values ​​under the dynamic room temperature setpoint within 24 hours, i.e., the predicted energy consumption value for the whole day, in kWh; For the hourly heat load prediction model in step S1, in actual use, The calculation needs to input all the variables required by the prediction model mentioned in step S1, among which the meteorological parameters are obtained from the weather forecast; For the average predicted percentage of dissatisfied rate of personnel during the working hours of a day, The PPD average number of working hours is taken as the optimization target, as calculated by formula (9), since the use time of this example is from 8:00 to 20:00, the thermal comfort during non-use time is not considered.

[0069] In this example, the decision variable determined in step S2 is the room temperature set value, and it is stipulated that the room temperature set value does not change within 3 hours, and the value range is {18℃, 19℃, 20℃, 21℃, 22℃, 23℃, 24℃}. The solving method is NSGA-II genetic algorithm.

[0070] The internal logic of the model predictive controller in step S3 of this example is as shown in Figure 3 .

[0071] The rolling optimization process of the model predictive controller in step S3 of this example is as shown in Figure 4 .

[0072] The internal logic of the model predictive controller in step S3 of this example adopts the enhanced elite reservation genetic algorithm (SEGA) to solve the rolling optimization process between the sampling time. The SEGA genetic algorithm is mainly used to solve the optimal hot water flow sequence at each sampling time: first, different hot water flow sequences are randomly generated, then they are input into the prediction model together with the real room temperature and meteorological forecast, and the room temperature prediction trajectory corresponding to the hot water flow sequence is obtained by repeatedly using the prediction model, then the room temperature prediction trajectory and the hot water flow are brought into the objective function to calculate the fitness, until the optimal hot water flow sequence and the room temperature prediction trajectory with the optimal cost are found, and finally the water pump frequency is adjusted to make the hot water flow through the radiator equal to the first item of the optimal hot water flow sequence, completing the solution at this time. The room temperature changes through the radiator of the heating room for a period of time, and at the next sampling time, the real room temperature is input into the prediction model, and the genetic algorithm is used to solve again, completing the rolling optimization.

[0073] The room temperature prediction model in this example selects a black box model, which meets the following functions: input current room temperature, current meteorological conditions and radiator hot water flow, and output indoor temperature at next time.

[0074] The target of the rolling optimization of the model predictive controller in this example is to track the room temperature set value obtained by optimization with the smallest room temperature fluctuation and water pump energy consumption, and the objective function is,

[0075] (11)

[0076] In the formula, N To predict the time domain, k =0 is defined as the sampling time, and each rolling optimization starts from... k =Starting from 0, y k For the first k Predicted room temperature values ​​at a time step y 0 represents the actual measured room temperature value. P k To predict the first in the time domain k Energy consumption of variable frequency water pump within a time step; The weighting coefficients are for room temperature fluctuations and variable frequency pump energy consumption. The prediction and control time domains are 2 hours, with a time step of 15 minutes, meaning the room temperature is adjusted every 15 minutes.

[0077] In this example, the rolling optimization of the model predictive controller uses the prediction of the hot water flow sequence in the time domain. U Each component is represented. u The optimization range is,

[0078] (12)

[0079] In the formula, The maximum flow rate allowed through the radiator is set to 0.15 kg / s in this example, with the optimization range upper limit set at 0.15 kg / s.

[0080] In this example, the constraints in the rolling optimization of the model predictive controller are as follows:

[0081] (13)

[0082] (14)

[0083] In the formula, The room temperature is a predicted value, and the room temperature at any given time should be greater than the 18°C ​​specified in the standard. Set the room temperature value for that moment. To allow for fluctuations in deviation, in this example .

[0084] In this example, the SEGA genetic algorithm is used to solve the rolling optimization of the model predictive controller. The tracking performance of model predictive control and traditional PID control for dynamic room temperature setpoints is shown in the attached figure. Figure 5As shown, the model predictive control has the ability of fast response and stabilizing the room temperature near the target value as soon as possible, and the room temperature under the model predictive control deviates from the set value to a smaller extent, so the model predictive control is more suitable for tracking the dynamic room temperature set value and has more advantages in the room temperature control of the heating room.

[0085] In this example, according to step S4, the user inputs the proportion of energy consumption and thermal comfort index according to his own preference into the SEGA optimization model to obtain the optimal room temperature set value trajectory T.

[0086] (15)

[0087] In the formula, are different objective functions, is the weight of the corresponding objective function set by the user;

[0088] In this example, according to step S4, the weights of the energy consumption objective function and the thermal comfort objective function in formula (15) are set to (1, 0), (0.75, 0.25), (0.5, 0.5), (0.25, 0.75) and (0, 1) respectively according to the user's preference, and the room temperature set value results of Dalian City from December 31 to January 2 of the next year in a typical year are obtained.

[0089] In this example, according to step S5, the model predictive controller is used to control the indoor temperature of the heating room to change according to the five different optimal room temperature set value trajectories T in example step S4.

[0090] Figure 6 Fig. 4 is a control effect diagram of the room temperature of this example under different energy consumption and thermal comfort ratios according to the present application.

[0091] Figure 7 Fig. 5 is a control effect diagram of this example under the same room temperature level as the present application under the traditional strategy.

[0092] Figure 8 Fig. 6 is a comparison diagram of the heating heat consumption of this example under the present application and the traditional strategy.

[0093] Figure 9 Fig. 7 is a comparison diagram of the heating pump consumption of this example under the present application and the traditional strategy.

[0094] By comparison, it is found that, compared with the traditional strategy, the heating heat consumption under the control of the application in five working conditions is reduced by 12.2%, 10.3%, 9.5%, 8.3% and 6.9% respectively, the greater the preference of the user to the energy saving target, the greater the reduction degree of the application to the heating heat consumption, and the more obvious the energy saving effect. It is calculated that, under the same indoor temperature level, the room temperature control method based on dynamic room temperature setting value and model predictive control proposed by the application is more energy-saving than the traditional strategy, and can reduce the total heating energy consumption including heating heat consumption and water pump conveying energy consumption by 9.7% on average.

[0095] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. It should be noted that, for those skilled in the art, without departing from the technical principles of the present application, a number of improvements and modifications can be made, and these improvements and modifications should also be considered as the protection scope of the present application.

Claims

1. A room temperature control method based on dynamic room temperature setpoint and model predictive control, characterized in that, The steps are as follows: Step S1: Determine the building or room to be heated, and establish an hourly heat load prediction model for the building or room to be heated. The hourly heat load prediction model uses a neural network prediction model. The training data is obtained through energy consumption simulation software or by collecting actual operating data. In order to improve the sensitivity of the hourly heat load prediction model to the room temperature setpoint, the room temperature setpoint in the training data needs to vary between 18 and 24°C. The input variable for the hourly heat load prediction model is: indoor design temperature; Is the current time a working day? Outdoor dry-bulb temperature; solar radiation levels. Staffing rate; Lighting utilization rate; Heat load values ​​for the previous 1 hour, 2 hours, and 24 hours at the current moment; The output variable of the hourly heat load prediction model is the hourly heat load value at the next moment; Step S2: Use the hourly heat load prediction model and the simplified PPD calculation program to construct a multi-objective optimization model, and determine the solution method; The reasonable simplification method for variables in the PPD calculation program is as follows: the mechanical work performed by the human body. =0, basic thermal resistance of clothing It is 1.4 clon, and the human metabolic rate is... 1 Met, mean radiation temperature The indoor air humidity is 2°C lower than the room temperature. Given the water vapor partial pressure at room temperature Tn and relative humidity of 40%, and an indoor air velocity v of 0.07 m / s; the simplified PPD calculation procedure is denoted as: ； The hourly heat load prediction model and the simplified PPD calculation program constitute a multi-objective optimization model. ； In the formula, This refers to the room's heat load index. Indicators for room thermal comfort; It is the sum of the predicted heat load values ​​under the dynamic room temperature setpoint within 24 hours, i.e., the predicted energy consumption value for the whole day, in kWh; The hourly heat load prediction model in step S1, in actual use, The calculation requires inputting all the variables required by the hourly heat load prediction model mentioned in step S1, where the meteorological parameters are obtained from weather forecasts; The average percentage of people dissatisfied with the forecast over a day. Depend on The calculation was obtained by the program; The decision variable for the multi-objective optimization model is the room temperature setpoint, and it is stipulated that the room temperature setpoint will not change within 3 hours, with a value range of {18℃, 19℃, 20℃, 21℃, 22℃, 23℃, 24℃}. The solution method is a genetic algorithm. Step S3: Design and add model prediction controllers for heated buildings or heated rooms; Add a model predictive controller for each building or for each heat user, depending on the heat user's needs; The prediction model can be a white-box model, a black-box model, or a gray-box model, but it must meet the following functions: input the current room temperature, current weather conditions, and radiator hot water flow rate, and output the indoor temperature at the next moment; The objective of the rolling optimization of the model predictive controller is to obtain the room temperature setpoint by minimizing room temperature fluctuations and minimizing pump delivery energy consumption. , is represented as , ； In the formula, N To predict the time domain, k =0 is defined as the sampling time, and each rolling optimization starts from... k =Starting from 0, y k For the first k Predicted room temperature values ​​at a time step y 0 represents the actual measured room temperature value. P k To predict the first in the time domain k Energy consumption of variable frequency water pump within a time step; The weighting coefficients for room temperature fluctuations and variable frequency pump energy consumption; In the rolling optimization of the model predictive controller, the hot water flow sequence in the time domain is predicted. U Each component is represented. u The optimization range is: ； In the formula, This is the maximum flow rate allowed through the radiator; The constraints in the rolling optimization of the model predictive controller are: ； ； In the formula, y k For the first k The predicted room temperature value for each time step should be greater than the 18°C ​​specified in the standard at any given time. Set the room temperature value for that moment. To allow for fluctuations in deviation; The internal logic of the model predictive controller is a rolling optimization process between the solution of the genetic algorithm and the sampling time. The genetic algorithm is mainly used to solve for the optimal hot water flow sequence at each sampling time: First, different hot water flow sequences are randomly generated using the SEGA algorithm. Then, along with the actual room temperature value and weather forecast, they are input into the prediction model. After iterative use, the prediction model will obtain the room temperature prediction trajectory corresponding to the hot water flow sequence. Then, the room temperature prediction trajectory and hot water flow are substituted into the objective function in the rolling optimization to calculate the fitness until the optimal hot water flow sequence and the corresponding cost-optimal room temperature prediction trajectory are found. Finally, the water pump frequency is adjusted so that the hot water flow through the radiator is equal to the first term of the optimal hot water flow sequence, completing the solution for that time. After a period of time, the room temperature of the heated building or heated room changes due to heat dissipation from the radiator. At the next sampling time, the actual room temperature is input into the prediction model, and the genetic algorithm is used again to solve the problem, completing the rolling optimization. Step S4: Users input the proportions of energy consumption and thermal comfort indicators into the SEGA optimization model according to their preferences to obtain the optimal room temperature setpoint trajectory T. The user's preference for the two objective functions will ultimately determine the choice of decision variables. In order to derive the final decision variables according to user preferences, the objective functions are normalized and then weighted: ； In the formula, i represents 1 and 2. , These are energy consumption indicators and thermal comfort indicators, respectively. The weights of the corresponding objective function set by the user; Step S5: The model predictive controller controls the indoor temperature of the heated building or heated room to change according to the optimal room temperature setpoint trajectory T.

2. The room temperature control method based on dynamic room temperature setpoint and model predictive control according to claim 1, characterized in that, The room temperature control method needs to be calculated between 20:00 every night and 8:00 the next day. The calculation content is the trajectory T of the optimal room temperature setpoint during the period from 8:00 the next day to 20:00 the day after next.

Images